Re: panic: vm_fault: fault on nofault entry, addr: fffffe0007e8e000

On Fri, Aug 01, 2003 at 10:59:06AM -0400, Andrew Gallatin wrote:
> 
> Kris Kennaway writes:
>  > On Fri, Aug 01, 2003 at 09:00:44AM -0400, Andrew Gallatin wrote:
>  > 
>  > > The crashdump might actually be useful here.  You'd have only the
>  > > trap() and vm_fault() frames, but at least you'd have information
>  > > about the state of the vm system.
>  > 
>  > Two crashdumps coming up!  I'll move them onto beast:/j/kris/crash
>  > together with the kernel.debug.
>  > 
> 
> I may have wasted your time.  The first one is unusable (lots of ddb
> cruft).  Damned gdb -k.  Grrr.

That one was the third panic I posted, which was the:

    trap entry = 0x2 (memory management fault)

> I don't have read perms on vmcore.{1,2}, so I don't know if they are
> helpful.   

Oops, sorry.

> If you're willing to get your traces via ddb's debug.trace_on_panic
> and to set debug.debugger_on_panic=0, then we might get at least a
> partial trace.  FWIW, I have to do this to get any sort of crashdump
> at all on SMP x86.  I'm amazed you were able to call doadump from ddb.
> When I try that on x86,   I just get a continuous stream of panics or
> fatal traps.

Hmm, I don't see this on the UP i386 machines.  Bento hasn't panicked
in a while (time to upgrade!), although I do seem to recall problems
the last time it did.

kris